How to Train NLTK NE CHUNKER

Agin das M

unread,

Dec 7, 2017, 7:21:26 AM12/7/17

to nltk-users

Hi,
I would like to train nltk ne chunker?
how can I do that ??

Denzil Correa

unread,

Dec 7, 2017, 7:51:56 AM12/7/17

to nltk-users

Did you try something out yourself? There are many simple links out there regardless including the NLTK book

https://streamhacker.com/2008/12/29/how-to-train-a-nltk-chunker/

http://www.nltk.org/book/ch07.html

--Regards,
Denzil
http://correa.in

--
You received this message because you are subscribed to the Google Groups "nltk-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nltk-users+...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Dimitriadis, A. (Alexis)

unread,

Dec 7, 2017, 7:57:14 AM12/7/17

to nltk-...@googlegroups.com

Hi Agin,

Take a look at the nltk book, free at http://nltk.org/book. It explains in detail how to train an NP chunker. For a named entity chunker you follow the same process, just with a corpus that annotates named entities rather than NPs.

Alexis

Jay Pratap Pandey

unread,

Dec 15, 2017, 5:35:47 PM12/15/17

to nltk-users

Hi

I read this book http://www.nltk.org/book/ch07.html

but i am still unable to trained the ne_chunk for Indian names

How can i solve this?

Denzil Correa

unread,

Dec 16, 2017, 4:55:35 AM12/16/17

to nltk-users

It will help if you specifically post (with code) what you did.

--Regards,
Denzil
http://correa.in

Jay Pratap Pandey

unread,

Dec 18, 2017, 6:58:52 AM12/18/17

to nltk-users

import nltk

from textblob.classifiers import NaiveBayesClassifier as NBC

from textblob import TextBlob

infile = open('Indian_names.csv', 'r')

training_data = infile.read()

infile1= open('Indian_names.csv', 'r')

testing_data infile1.read()

model = NBC(training_data)

print("Accuracy:", model.accuracy(test_data))

Dimitriadis, A. (Alexis)

unread,

Dec 18, 2017, 7:21:08 AM12/18/17

to nltk-...@googlegroups.com

This code uses textblob, not the nltk. Textblob provides wrappers around various nltk functions, but you cannot expect nltk users to know how to use it. I don’t.

The textblob code you included has nothing to do with the chapter of the nltk book that you say you read, or with named entity recognition for that matter. I recommend actually reading chapter 7 of the nltk book, and some prior chapters to help you understand it. It’s all laid out pretty clearly.

However, be warned that you can’t do named entity recognition with a list of names; you need an annotated corpus. Otherwise you’re just doing trivial word lookup.

Alexis

PS. It’s a terrible idea to “test” with the same data you trained with.

Reply all

Reply to author

Forward